Identify the CPU that is handling a session
Hi community,
I have a FGT VM with one CPU (of 4) that reaches 100% peaks, while the others are idle. Due to the characteristics of the traffic I believe this is "normal" (there is a fixed session between 2 IPSec hosts that always use the same source IP/port and the same destination IP/port with over 800Mbps throughput, which is handled by a single CPU, while the other ones share the rest of the load). I believe this CPU is dropping packets because it can't handle the total amount of traffic, and I would like to demonstrate this to the customer, but how?
I can easily:
- identify the single session (knowing its characteristics of src and dst) with 'diag sys session list'
- show that one CPU is overloaded with 'diagnose sys performance status',
but how do I relate the session from the first command to the CPU status from the second command? How do I show that there is packet loss?
Could you suggest some commands that highlight these aspects?
thanks guys,
Angelo
